This invention relates to compression/decompression device (codec) speed selection within a voice packet network, and more specifically to method and apparatus for dynamic selection of codec speed based upon demand to provide an optimal match between voice and data payload bandwidth requirements and codec bandwidth capacities.
In voice packet networks such as voice over frame relay (VoFR) or voice over Internet protocol (VoIP), voice overlays data in digitized, packetized, usually compressed form. Thus, relatively low-speed voices and relatively high-speed modem data share codecs used to compress and decompress the data at the endpoints. If it is possible to determine the voice or modem nature of the payload during placement of a telephone call, then it is possible to allocate an appropriate bandwidth of an appropriate codec to the telephone call. However, payload demands may vary even within a given telephone call. For example, a telephone line may have a handset as well as a fax or modem and a telephone call may switch between data and voice during a given telephone connection. As a result, even if the payload requirement is properly understood at the beginning of a telephone call, it may change its nature during the telephone call. This can result in loss of data, if bandwidth is under-allocated. It can also result in waste of channel resource, if bandwidth is over-allocated. There is presently no mechanism for productively handling the dynamically changing demands of channel bandwidth in such mixed-use network contexts.
Dynamic compressor/decompressor (codec) device speed selection includes monitoring a call in a telephone system to determine the type of codec being used thereon and to measure the payload on the codec. The capacity of the codec is determined based upon the determined type of codec being used. Speed selection further includes comparing the determined capacity with the measured payload; and selectively switching the call to a higher operational speed based upon the result of the comparing. Preferably, the monitoring includes detecting a modem answer tone of the telephone call and basing the determination of the type of codec at least partly thereon. The monitoring also preferably includes detecting silence of the telephone call and basing the selective switching at least partly thereon. Echo cancellation and/or suppression is disabled upon such selective switching. In accordance with one embodiment of the invention, the switching includes requesting an up-speed codec. In such case, the method further includes verifying network bandwidth prior to the requesting of the up-speed codec. The network bandwidth is pre-allocated in accordance with the invention after the up-speed codec is requested and prior to the selective switching of the call to a higher operational speed. Finally, the monitoring further includes determining whether the call is silent and upon such determination automatically switching the call back to a lower operational speed.